Novel adaptive UKF and its application in integrated navigation
HU Gao-g · Journal of Chinese Inertial Technology · 2014
This paper presents a novel adaptive UKF with noise statistic estimator for the purpose of improving the filtering accuracy of integrated navigation systems. The covariance matching technique is employed in the proposed algorithm, and the innovation and residual sequences are used to estimate and adjust the covariance matrices of the process and measurement noises online. The proposed algorithm enhances the adaptive capability of the UKF and overcomes the limitation of the standard UKF, otherwise the filtering solution will be deteriorated or even divergent as the system noise statistics are unknown or inaccurate. The proposed algorithm is applied to the SINS/BDS integrated system for simulation in comparison with the standard UKF and robust UKF. The simulation results demonstrate that the horizontal position error and vertical error obtained by the proposed adaptive UKF are within [-6.2 m, +6.4 m] and [-9.8 m, +8.6 m], respectively. The performance of the proposed algorithm is significantly superior to that of the standard UKF and robust UKF, leading to improved calculation precision of the integrated navigation system.